Model Challenger vs Benchmark — Bias Analysis — stdsquare²

Model vs benchmark — deteksi bias sistematis

Bias rata-rata (mean error %)0,5

Volatilitas error σ_e (%)1,2

Autokorelasi error (ρ lag-1)0,3

Sample size T250

RMSE

—

akar MSE

MAE

—

mean abs error

Diebold-Mariano

—

vs benchmark

Status

—

model/benchmark

Error model vs aktual — pola sistematis